A early mode decision algorithm based on statistics and machine learning for enhancement layers in H.264 Scalable Video Coding

In this paper, a early mode decision algorithm is proposed to reduce the complexity of the mode selection process for enhancement layers in H.264 Scalable Video Coding. Generally, the proposed algorithm consists of the following three main steps, which are applied to the enhancement layer. Firstly, we divide all the macroblocks into 4 classes according to the mode of collocated macroblocks in the base layer. Then, according to the mode of neighboring macroblocks, the macroblocks are subdivided with trained BP (Back Propagation) network. At last, for different cases we choose different mode selection algorithms. Compared to JSVM 9.18, experiment results show that, with this algorithm, 30% encoding time can be saved with a negligible loss in BDSNR, and BDBR can be significantly reduced.